Similar to my previous article on how to configure Jupyter to work with R, I want to do some data processing using NodeJS and Dataframes.

Step 0 - Prerequisite

Learn how to install any version of node using node version manager

nvm install 12.0.0

Pkg-Config

brew install pkg-config

ZeroMQ

brew install zeromq

Do not use zmq. zmq is deprecated.

iPython

brew install ipython

Step 1 - Clone from Github

Change directory to your Mac Desktop.

mkdir ~/Documents/MY_LIBS/
cd ~/Documents/MY_LIBS/

Download the NodeJS Kernel from Github.

git clone https://github.com/notablemind/jupyter-nodejs.git && \
cd jupyter-nodejs/

Step 2 - Create a Directory

The command below assumes that you already have Jupyter installed and you're only adding a new directory within ~/.ipython/ for the remaining steps.

mkdir -p ~/.ipython/kernels/nodejs/

Step 3 - Update and Install

Make sure you are using the latest version of node-gyp

npm i npm -g

This module is compatible with ZeroMQ versions 2, 3 and 4.

npm i --save zeromq@^4.0.0 

Anything above 4.0 may not yet work.

Then install the package

npm i && node install.js

Step 4 - Build

Let's do some Node Package Manager building and extension building.

npm run build
npm run build-ext

Step 5 - Double-Check

If you can run this command, then you're right.

jupyter console --kernel nodejs

jupyter-nodejs-cli

Step 6 - Done

Here's what it will look like if you open your notebook within a browser.

jupyter-nodejs


Troubleshooting

The most common error I see has to do with the ZeroMQ package.

Resources